"""Benchmark fixtures for hypergraph performance tests."""
import random
import pytest
from hypabase.engine import Hyperedge, HypergraphStore, Incidence, Node
def generate_random_graph(
num_nodes: int,
num_edges: int,
avg_cardinality: float = 2.5,
directed_ratio: float = 0.3,
seed: int = 42,
) -> HypergraphStore:
"""Generate a random hypergraph for benchmarking.
Args:
num_nodes: Number of nodes to create
num_edges: Number of hyperedges to create
avg_cardinality: Average number of nodes per edge
directed_ratio: Fraction of edges that are directed
seed: Random seed for reproducibility
Returns:
HypergraphStore with random data
"""
rng = random.Random(seed)
store = HypergraphStore()
# Create nodes
node_types = ["table", "column", "concept", "metric"]
node_ids = [f"node_{i}" for i in range(num_nodes)]
for node_id in node_ids:
store.add_node(
Node(
id=node_id,
type=rng.choice(node_types),
properties={"index": int(node_id.split("_")[1])},
)
)
# Create edges
edge_types = ["foreign_key", "concept_mapping", "synonym", "hierarchy"]
for i in range(num_edges):
# Determine cardinality (2 to avg*2, centered around avg)
cardinality = max(2, int(rng.gauss(avg_cardinality, avg_cardinality / 2)))
cardinality = min(cardinality, num_nodes)
# Select random nodes
selected_nodes = rng.sample(node_ids, cardinality)
# Determine if directed
is_directed = rng.random() < directed_ratio
# Create incidences
incidences = []
for j, node_id in enumerate(selected_nodes):
if is_directed:
# First half are tails, second half are heads
direction = "tail" if j < len(selected_nodes) // 2 else "head"
else:
direction = None
incidences.append(Incidence(node_id, direction=direction))
store.add_edge(
Hyperedge(
id=f"edge_{i}",
type=rng.choice(edge_types),
incidences=incidences,
confidence=rng.random(),
)
)
return store
@pytest.fixture
def graph_1k() -> HypergraphStore:
"""1K nodes, 5K edges - small benchmark graph."""
return generate_random_graph(
num_nodes=1000,
num_edges=5000,
avg_cardinality=2.5,
seed=42,
)
@pytest.fixture
def graph_10k() -> HypergraphStore:
"""10K nodes, 50K edges - medium benchmark graph."""
return generate_random_graph(
num_nodes=10000,
num_edges=50000,
avg_cardinality=2.5,
seed=42,
)
@pytest.fixture
def graph_100k() -> HypergraphStore:
"""100K nodes, 500K edges - large benchmark graph."""
return generate_random_graph(
num_nodes=100000,
num_edges=500000,
avg_cardinality=2.5,
seed=42,
)
@pytest.fixture
def sparse_graph_10k() -> HypergraphStore:
"""Sparse 10K graph - fewer edges, lower connectivity."""
return generate_random_graph(
num_nodes=10000,
num_edges=10000, # 1 edge per node on average
avg_cardinality=2.0,
seed=42,
)
@pytest.fixture
def dense_graph_10k() -> HypergraphStore:
"""Dense 10K graph - more edges, higher connectivity."""
return generate_random_graph(
num_nodes=10000,
num_edges=100000, # 10 edges per node on average
avg_cardinality=3.0,
seed=42,
)
@pytest.fixture
def hyperedge_heavy_10k() -> HypergraphStore:
"""10K graph with many large hyperedges."""
return generate_random_graph(
num_nodes=10000,
num_edges=20000,
avg_cardinality=5.0, # Larger edges
seed=42,
)
